Greening the cloud: Digital infrastructure for a net-zero world

06 October 2023

Anne Benitez Cannelles, Analyst

Please note: This article is designed to be thought leadership content to offer big picture views and analysis of interesting issues and trends that matter to our clients and the world in which we live. It is not designed to be taken as expert advice, investment advice or a recommendation and any reference to specific companies is therefore not an opinion as to their present or future value or broader ESG credentials. Reliance upon any of the information in this article is at the sole discretion of the reader. Some of the views and issues discussed in this article may derive from third-party research or data which is relied upon by Barclays Private Bank and may not have been validated. Such research and data are made available as additional information for the reader where appropriate.

The exponential leap in digital adoption following the pandemic has escalated concerns about the environmental footprint of information and communications technology (ICT) infrastructure. How can we meet growing digital demand while improving sustainability?

Data centres are the infrastructure behind everything we consume online. These groups of networked computers range from 1000sq metre boxes – often found in remote corners of the world and which ‘ground’ the cloud – to highly-customised units built into other complexes. Whatever their design, they respond to almost everything we choose to do on the internet and store, distribute and process most information dear to us. However, their power intensity attracts scrutiny from environmentalists and policy makers alike, and is bringing into question their role in a low-carbon future. 

The International Energy Agency (IEA) estimated that data centre electricity use in 2021 was around 1% of global electricity demand1. The electricity used to meet these power demands still largely relies on fossil fuels and creates significant pressures on local energy grids. There is also the issue of keeping them cool: servers running within data centres generate high levels of heat and the cooling can make up to 40% of the energy bill2. Temperatures are often brought down by evaporating millions of gallons of water, which can significantly strain local resources. 

The IT industry is moving fast with solutions, such as harnessing excess heat for use in other buildings, or improving airflow management, whereby fan speeds are adapted to create ideal circulation channels. However, as of 2021, Nature Research has estimated that a medium-sized data centre uses as much water as three average hospitals3

Cutting emissions through scale

Much of the technological progress helping to further a better understanding of the world, and contributing to the fight against climate change, is sustained through infrastructure at the scale of large data centres. For example, machine learning is powered by a series of algorithms (some of them aptly called neural networks) that mimic the way the human brain operates. The workloads involved are often so immense that up until the last decade, it was impractical to run them on traditional computers. Some of the biggest breakthroughs of the past decade have come from this leap in computing capacity. For example, big data and AI are helping to map out malaria outbreaks, deploy renewable energy grids, and explore protein folding at a depth and range unprecedented a decade ago. 

In the coming decades, artificial intelligence applications will likely support better climate change modelling and research, design more energy-efficient buildings and improve our understanding of battery systems and storage capacities, amongst other things. For example, the US National Oceanic and Atmospheric Administration is already using AI to enhance its use of satellites and environmental data to improve extreme weather forecasting4

Most importantly, large-scale data centres embody the fundamentals of digital innovation: the potential for economies of scale. Since 2010, the shift away from small, inefficient, on-premise servers to efficient, hyper-scale data centres has seen the supply of digital services outpace the growth of their carbon footprint5. So much so that between 2015-2019, while the global energy use for data centres remained flat, internet traffic tripled6.

Another indicator of improving efficiency is the trend in Power Usage Effectiveness, or PUE, in recent years. This standard industry metric measures the energy efficiency of data centres, where the lower the value, the less energy is being used. While in the past three years we’ve seen a slight stalling of this metric, over the last decade, we’ve seen the average annual PUE for large data centres come down from 2.5 to 1.55 – with some even reaching 1.2 PUE. By staying ahead of the innovation curve, hyper-scalers are estimated to dominate 61% of all installed data centre servers7

At these scales, marginal efficiency improvements in their design can lead to disproportionate emissions savings per user. Instead of hundreds of hardware systems that would typically characterise an individual enterprise, large providers are able to run a very small number of configurations that make it easier to address bottlenecks. For example, the energy efficiency of any data centre is largely a function of the intensity of their computational and cooling sub-systems. Optimisations in both have been largely achieved by making the transition away from monolithic, customised data centres accommodating individual demands, to more industrial, prefabricated modular buildings that have allowed companies to connect built-in factory models designed for better ‘behaviour’. 

Connected networks 

Technology sourcing policies are also being revisited, with the goal of better aligning the number of servers with the demand. Ensuring reliability is a primary concern for data centre operators, which leads to the procurement of “just in case” equipment to back up the primary servers in case of failure. Now, however, resilience is starting to be provided at an application level rather than facility level, by making use of connected networks, where a number of independent sites use a shared internet connection or VPN. For example, if a given server experiences downtime, the application running in it will work its way around other connected servers. 

The industry is even venturing into the idea of ‘green coding’. While still largely in the realm of the academic world, this is all about designing software that minimises the energy requirements of the hardware running it8. Hyper-scale operators are also unlocking large-scale purchases of renewable power, which would be impossible in a highly distributed system. The promise of consistency and volume from these operators often makes otherwise unviable renewable projects possible. And in cases where the original anchor customer relocates or falls through, standardisation of contracts is becoming more common to de-risk long-term agreements. 

Reason for optimism

The global data load is expected to rise to a staggering 221 zettabytes by 20269. The radical reduction in resource intensity and procurement of green energy therefore have implications for the design and operation of data centres at all stages of their lifecycle. In this way, the industry’s big players are demonstrating that it is possible to future-proof the industry. 

By debunking the assumption that doing or producing less is the only way to reduce environmental impacts, we are seeing economies of scale deliver on the demands of the digital age without straying from a solid path towards sustainability. More work is required, but there is reason for optimism.

Related articles


This communication is general in nature and provided for information/educational purposes only. It does not take into account any specific investment objectives, the financial situation or particular needs of any particular person. It not intended for distribution, publication, or use in any jurisdiction where such distribution, publication, or use would be unlawful, nor is it aimed at any person or entity to whom it would be unlawful for them to access.

This communication has been prepared by Barclays Private Bank (Barclays) and references to Barclays includes any entity within the Barclays group of companies.

This communication: 

(i) is not research nor a product of the Barclays Research department. Any views expressed in these materials may differ from those of the Barclays Research department. All opinions and estimates are given as of the date of the materials and are subject to change. Barclays is not obliged to inform recipients of these materials of any change to such opinions or estimates;

(ii) is not an offer, an invitation or a recommendation to enter into any product or service and does not constitute a solicitation to buy or sell securities, investment advice or a personal recommendation; 

(iii) is confidential and no part may be reproduced, distributed or transmitted without the prior written permission of Barclays; and

(iv) has not been reviewed or approved by any regulatory authority.

Any past or simulated past performance including back-testing, modelling or scenario analysis, or future projections contained in this communication is no indication as to future performance. No representation is made as to the accuracy of the assumptions made in this communication, or completeness of, any modelling, scenario analysis or back-testing. The value of any investment may also fluctuate as a result of market changes.

Where information in this communication has been obtained from third party sources, we believe those sources to be reliable but we do not guarantee the information’s accuracy and you should note that it may be incomplete or condensed.

Neither Barclays nor any of its directors, officers, employees, representatives or agents, accepts any liability whatsoever for any direct, indirect or consequential losses (in contract, tort or otherwise) arising from the use of this communication or its contents or reliance on the information contained herein, except to the extent this would be prohibited by law or regulation.